Lactiplantibacillus plantarum P9 for chronic diarrhea in young adults: a large double-blind, randomized, placebo-controlled trial

Chronic diarrhea is a common functional gastrointestinal disorder affecting an estimated 3–20% of the global population, characterized by more than three bowel movements or loose stools per day for at least four weeks. Standard therapies (e.g., loperamide, diphenoxylate, clonidine, opioids, anticholinergics) can cause side effects and raise concerns about resistance and persistence, making management challenging. Individuals with diarrhea often exhibit reduced gut microbial diversity, altered microbial metabolites (bile acids, short-chain fatty acids, amino acids), and emerging evidence implicates the gut phageome in gastrointestinal disorders. Probiotics may mitigate diarrhea by modulating inflammation, restoring gut microbiota balance, and enhancing barrier function, although prior human trials show mixed efficacy depending on strain, dose, sample size, and duration. Lactiplantibacillus plantarum P9 (P9), isolated from traditional acidic gruel, shows high tolerance to gastroenteric fluids and immunomodulatory, anti-inflammatory, and microbial homeostasis-modulatory activities, and has previously alleviated chronic constipation with favorable microbiome and metabolite shifts. The present large, randomized, double-blind, placebo-controlled clinical study aimed to evaluate the efficacy and safety of P9 for chronic diarrhea over 4 weeks, with the primary endpoint being the diarrhea symptom severity score and secondary endpoints including stool consistency, frequency, urgency, and multi-omics changes in the gut microbiome and metabolome.

Prior clinical studies of probiotics for chronic diarrhea and IBS-D demonstrate variable outcomes: several single- or multi-strain trials reported improvements in defecation frequency, stool consistency, and quality of life, while others showed no significant benefit after probiotic fermented milk interventions. The effects appear strain-, dose-, cohort-, and duration-dependent. Nonclinical evidence indicates probiotics can reduce diarrhea by modulating gut motility and microbiota composition (e.g., Saccharomyces boulardii in murine models; multispecies probiotics in calves). P9 specifically has documented tolerance to GI conditions, immunomodulatory properties, and the capacity to modulate intestinal microbial communities and metabolites; in a prior RCT it improved constipation with increases in functional taxa and metabolites. Additionally, SCFA-producing taxa and bile acid metabolism are implicated in diarrhea pathophysiology, and phageome alterations have been associated with gastrointestinal diseases including IBD and colorectal cancer, supporting a rationale for multi-omics assessment during probiotic intervention.

Design: Randomized, double-blind, placebo-controlled clinical trial conducted from October 2020 to May 2021; registered at ChiCTR2000038410 and approved by the Ethics Committee of the First Affiliated Hospital of Nanchang University (No. IIT [2020] Clinical Ethics Review No. 002). Following a two-week run-in, participants were randomized 1:1 to P9 or placebo using computer-generated codes by independent administrators; allocation was concealed and investigators, clinicians, and participants were blinded. Intervention: P9 powder containing 1.0 × 10^10 CFU in 2 g taken once daily 30 minutes after meals, mixed with warm water, for 28 days (day 0–28) followed by a 14-day post-intervention follow-up (to day 42). Placebo: maltodextrin powder identical in appearance. Compliance was assessed by counting returned doses. Participants: 189 adults (18–65 years) with chronic diarrhea per Rome IV criteria (loose/watery stools > 25% of bowel movements in last 3 months; Bristol types 5–7; symptoms onset ≥ 6 months), normal routine fecal tests (ages 18–50), and without organic GI disease, IBD, celiac disease, or other exclusion conditions (e.g., recent antibiotics/probiotics, pregnancy/lactation, severe comorbidity, psychotropic drugs, long-term antidiarrheals). ITT population n=189 (P9 n=93; placebo n=96). PP analysis included n=170 (85 per group) after exclusions (COVID-19 infection, antibiotics use, refusals/missing diaries). Primary endpoint: Diarrhea symptom severity score derived from three GSRS items—(1) increased stool passage (0–3), (2) loose stools (Bristol-guided, 0–3), and (3) urgency (0–3); summed as weekly averages. Secondary endpoints: stool consistency (Bristol 1–7), bowel movement number, fecal urgency score (0–3), DASS-21, and fecal multi-omics (metagenome, phageome, untargeted and targeted metabolomics). Data collection: Daily electronic stool diary; outcomes summarized at days 14, 28, 42. Biospecimens: Fecal samples collected at days 0, 28, 42 with DNA preservative and stored at −80°C. Metagenomics: PP population samples from 169 subjects (pro n=84, pla n=85) sequenced on Illumina NovaSeq 6000; assembly with MEGAHIT; binning with MetaBAT2, VAMB, DAS Tool; genome quality by CheckM; dereplication with dRep to define 629 species-level genome bins (SGBs); taxonomic annotation by Kraken2; abundance by CoverM. Functional reconstruction: GMMs predicted via MetaCyc/KEGG; modules identified with Omixer-RPM (c≥0.66); MelonnPan predicted bioactive metabolites; CAZymes annotated by dbCAN2. Phageome: Viral contigs quality by CheckV; identification by VIBRANT; clustering to vOTUs (≥5 kb; 95% identity over 80% length) by CD-HIT; compared to Metagenomic Gut Virus catalog; abundances by CoverM. Untargeted metabolomics: LC-MS on PP fecal samples (n=161; pro n=80, pla n=81); data processed by ProteoWizard and XCMS; PCA and PLS-DA, with significant metabolites selected by VIP>2, Wilcoxon P<0.05. Targeted metabolomics: Quantification of SCFAs (GC-MS) and bile acids (LC-MS) on PP samples with sufficient volume (n=127; pro n=66, pla n=61). Safety: Adverse events recorded (systemic infection, deleterious metabolic activities, immune stimulation, GI side effects). Statistics: R (v4.1.0); chi-square for categorical variables; Wilcoxon rank-sum (paired/unpaired as appropriate) for outcomes; ITT imputation via baseline-observation carried-forward; non-parametric covariance analysis (sm package) using baseline as covariate; alpha/beta diversity by Shannon/Simpson and PCoA with adonis; Procrustes analysis; Pearson correlations; significance at P<0.05.

Participants: 215 assessed; 26 excluded; 189 randomized (P9 n=93, placebo n=96). After 4 weeks, PP included 170 (85 per group). Baseline: groups balanced in age (mean ~22 years), sex, BMI, ethnicity, allergies, smoking, comorbidities, and prior medications. Primary outcome: Diarrhea symptom severity score decreased more with P9 than placebo—ITT: 20.0% reduction (mean [SD] at day 28: P9 1.2 [0.8] vs placebo 1.5 [0.9]; P=0.050); PP: 21.4% reduction (P9 1.1 [0.7] vs placebo 1.4 [0.8]; P=0.048). Non-parametric covariance analysis confirmed between-group differences at days 14 (P=0.003), 28 (P<0.001), 42 (P=0.011). Secondary clinical outcomes: Stool consistency (Bristol) lower in PP at day 28 with P9 (4.7 [0.7] vs 4.9 [0.6]; P=0.050); not significant in ITT (P=0.052). No significant between-group differences in bowel frequency, fecal urgency, depression, anxiety, or stress (P>0.05). Safety/compliance: No serious adverse events; fewer AEs in P9 vs placebo at day 28 (7 vs 15) and day 42 (4 vs 5); high compliance (mean ~102–104%). Microbiome (bacteria): No alpha-diversity differences (Shannon/Simpson, P>0.05); beta-diversity differed at days 28 (R^2=0.015, P=0.012) and 42 (R^2=0.014, P=0.018). Of 629 SGBs, 29 were differentially abundant during/after intervention. Day 28: higher Lactiplantibacillus plantarum and Ruminococcus A faecicola; lower Mediterraneibacter torques, Eubacterium I ramulus, Enterocloster sp000431375 (P<0.05). Day 42: higher Butyricicoccus A sp002395695, Streptococcus thermophilus; lower Phascolarctobacterium faecium and Faecalibacterium sp. Eight SGBs differed consistently at both days 28 and 42 (e.g., more Acutalibacteraceae sp000431775, Paraprevotella xylaniphila; fewer Coprococcus sp., Butyricimonas virosa). Correlations: Stool consistency positively correlated with Eubacterium I ramulus (r=0.2, P=0.019); stress positively with Enterocloster sp000431375 (r=0.2, P=0.008); bowel frequency positively with Mediterraneibacter torques (r=0.2, P=0.002); anxiety negatively with Ruminococcus A faecicola (r≈−0.2). Phageome: 94,384 non-redundant vOTUs (41,059 assigned to 13 families; predominance of Siphoviridae, Myoviridae, Microviridae, Podoviridae, crAssphage; 87.7% Caudovirales). No significant beta- or alpha-diversity differences overall; Simpson index trend at day 28 (P=0.056). Strong positive bacteria–phage diversity correlation (Shannon R=0.928, P=9.21e−21) and Procrustes correlation (R=0.818, P=0.001). Fewer Podoviridae and Myoviridae sequences with P9 at day 42 (P<0.05). Family-level weak positive correlations with symptoms: Microviridae and crAssphage with fecal urgency (r≈0.2), Herelleviridae with bowel frequency (r=0.2). Some specific vOTUs showed stronger correlations with urgency (e.g., Microviridae vOTU-39925, vOTU-59687; several crAssphage vOTUs). Predicted functions/metabolites: 72 GMMs across 11 categories; Bacteroidales, Lachnospirales, Oscillospirales encoded diverse modules. Predicted bioactive metabolites differed at day 28 (R^2=0.028, P=0.013) with 18 P9-responsive metabolites: enriched cholate, chenodeoxycholate, C16 carnitine, creatine, bilirubin; depleted deoxycholate and glutamate (P<0.05). CAZymes: 26,170 genes identified (GH 14,598; GT 6,108; CE 2,726; CBM 1,690; PL 810; AA 238); 15 subfamilies enriched with P9 at day 28/42 (e.g., GH108, GH13 variants, GH158, GH64, GH5_18; CBM4, CBM56; GT21, GT60, GT74; PL13, PL10_2; CE16). Untargeted fecal metabolomics: 21 differential metabolites at days 28/42; notable increases in caffeic acid, taurine, 3-hydroxypentanoic acid, hexacosanoic acid, cerotic acid (P<0.05). 3-hydroxypentanoic acid correlated negatively with stool consistency (r=−0.22, P=0.048) and diarrhea severity (r=−0.23, P=0.039). Targeted metabolomics: Higher acetic acid and butyric acid with P9 at day 28 (P<0.05); no differences in isovaleric, valeric, propionic, isobutyric acids. Lower deoxycholic acid (DCA) with P9 (P<0.05); no differences in cholic acid (CA) and chenodeoxycholic acid (CDCA).

The trial demonstrates that P9 administration modestly but significantly reduces diarrhea symptom severity compared with placebo in both ITT and PP analyses, with a small improvement in stool consistency in PP analysis. While both groups improved over time, the greater reduction in the P9 arm suggests a true treatment effect beyond placebo. Multi-omics data support plausible mechanisms: P9 shifted gut microbiota structure without altering alpha diversity, enriching taxa linked to SCFA production and gut homeostasis (e.g., Lactiplantibacillus plantarum, Ruminococcus A faecicola, Paraprevotella xylaniphila, Butyricicoccus A sp.) and reducing taxa associated with IBS-D or dysbiosis (e.g., Mediterraneibacter torques, Coprococcus sp., Enterocloster sp.). These microbial shifts coincided with enrichment of CAZyme subfamilies and predicted SCFA-related functional modules, and were corroborated by increased fecal acetate and butyrate. P9 also modulated bile acid metabolism, lowering deoxycholic acid—a secondary bile acid implicated in secretory diarrhea and mucosal permeability—while predicted profiles indicated higher cholate and chenodeoxycholate and lower deoxycholate. Amino acid-related metabolites changed as well, with decreased glutamate and increased taurine, aligning with potential improvements in motility, barrier integrity, and inflammation. The phageome exhibited strong covariation with bacterial diversity, selective reductions in Podoviridae and Myoviridae, and weak family-level associations with symptoms; specific vOTUs correlated more strongly with fecal urgency, suggesting that finer-resolution viral dynamics may contribute to symptom modulation. Overall, P9 appears to alleviate diarrhea via multi-level modulation of gut ecology and metabolites that support barrier function and reduce pro-secretory/pro-inflammatory signals.

In young adults with chronic diarrhea, four weeks of Lactiplantibacillus plantarum P9 resulted in a modest but statistically significant improvement in diarrhea symptom severity compared with placebo, with supportive shifts in gut microbiota composition, functional enzyme capacity, and metabolites, including increased acetate and butyrate and reduced deoxycholic acid. The intervention was well tolerated with no serious adverse events. These findings support probiotics as a complementary strategy for managing chronic diarrhea. Future work should employ more stringent diagnostic workups (including blood biomarkers, inflammatory indices, imaging/endoscopy), longer intervention and follow-up periods, larger and more phenotypically stratified cohorts, comprehensive dietary assessment and control, and higher-resolution virome analyses to clarify phage contributions. Mechanistic studies integrating host biomarkers (e.g., barrier and inflammatory markers) and planned miRNA analyses will further elucidate P9’s modes of action and identify responders.

Key limitations include: (1) potential misclassification and heterogeneity due to reliance on Rome IV criteria without comprehensive laboratory/endoscopic exclusion of IBS-D subtypes (pandemic limited access to tests); (2) absence of blood-based physiological, barrier, and inflammatory markers; (3) short 28-day intervention and 14-day follow-up, which may be insufficient for maximal probiotic effects; (4) evidence of placebo effects and marginal significance for some outcomes (e.g., stool consistency significant only in PP); (5) high inter-individual variability and possible presence of non-responders; (6) incomplete and non-quantitative dietary data; (7) reduced final enrollment (189 vs planned 200) though meeting minimum sample size requirements; (8) incomplete stool volumes leading to reduced sample sizes for targeted metabolomics; (9) limited resolution of virome analysis tools/databases, yielding weak family-level associations.